Structural and Magnetic Properties of the New SrLa₂FeCuSbO₉ Perovskite Phase

Abstract

We report on the synthesis and structural and magnetic properties of the new SrLa₂FeCuSbO₉ perovskite. The compound was prepared by the solid-state reaction method. The crystalline structure was studied by X-ray diffraction experiments using the Rietveld refinement analysis. SrLa₂FeCuSbO₉ crystallizes in a monoclinic system (P2₁/n space group). Its unit cell parameters are a = 5.6101(2) Å, b = 5.6125(2) Å, c = 7.9358(3) Å, and β = 90.094(2)°. Sr²⁺ and La³⁺, in SrLa₂FeCuSbO₉, are statistically distributed within the A-perovskite sites. The obtained cationic distribution can be illustrated by the [SrLa₂]_A[(SbCu_0.5)_2c(FeCu_0.5)_2d]_BO₉ crystallographic formula. SrLa₂FeCuSbO₉ was found to exhibit a paramagnetic–ferrimagnetic transition around T_C = 215 K. At high temperature, the magnetic susceptibility was described by the Curie–Weiss law. The experimental effective moments were explained by only spin contribution for Cu²⁺ and Fe³⁺ ions. At low temperature, the compound presents a ferrimagnetic state with signs of frustration.

1 Introduction

Perovskite compounds of the general formula ABO₃, where A is an (s or f) block cation, and B is a (d or p) block cation, have been widely studied because of their very interesting and surprising physical and chemical properties [1]. These compounds represent a large percentage of systems currently studied and continue to interest scientists. In fact, it is due to the remarkable flexibility of this structure which can accommodate a wide variety of chemical elements with different oxidation states, which gives a large number of compounds with different fundamental and interesting chemical and physical properties such as electronic structures ranging from insulating to metallic [2], superconductivity [3], ferromagnetic [4], ferroelectric [5], magnetoresistance [6], and multiferroic [7]. In ABO₃ perovskite phases, if two or three different types of atoms are introduced into A and/or B sites, more complex structures, such as double perovskite AA′BB′O₆ and triple perovskite A₂A′B₂B′O₉, can be adopted [8]. In recent years, A₃B₂B′O₉ triple perovskites have been of great interest, because they present new promising physical properties such as relaxor ferromagnetic [9]. Recently, SrLa₂FeCoSbO₉ compounds have been synthesized and their structural and magnetic properties have been studied [10]. It has been found that this compound presents a relative height magnetic transition to a ferrimagnetic phase (T_C = 215 K). This result was related to the presence of two different magnetic (B, B′) cations in the octahedral sites. Indeed, this compound adopts a monoclinic perovskite structure (P2₁/n space group) with two distinct octahedral sites; one of these sites is occupied by (2/3Co²⁺, 1/3Fe³⁺) and the other by (2/3Sb⁵⁺, 1/3Fe³⁺). Similar compounds such as La₂SrCo₂FeO₉ [11], La₂SrFe₂CoO₉ [12], and Sr₃CrFeMoO₉ [13] have been investigated. The earlier three compounds were shown to present a ferrimagnetic state below 123 K, 280 K, and 400 K, respectively. The high magnetic transition temperature observed in these compounds seems to be linked to the presence of Fe³⁺ ions which imposes strong magnetic interactions. Several results show that the compounds which contain the same magnetic cations in two octahedral sites present a relatively low magnetic transition temperature T_C such as the compounds La₃Ni₂SbO₉ (T_C = 105 K) [9], La₃Co₂SbO₉ (T_C = 55 K) [14], LaSr₂Cr₂SbO₉ (T_C = 150 K) [15], SrLa₂Ni₂TeO₉ (spin glass below 35 K) [16], and Sr₃Fe₂TeO₉ (spin glass below 80 K) [17]. It should be noted that in this type of compounds the crystal structure, the nature of the chemical elements which are in A, B, and B′ sites, and the cation distribution between the two B and B′ sites of the perovskite structure play a determinant role on their magnetic properties. Indeed, these parameters can modulate the competition between the different types of magnetic interactions which lead to particular effects, such as magnetic frustration or high Tc ferromagnetic properties. In our previous structural characterization study, the structure of the [CaLa₂]_A[CaTi₂]_BO₉ phase (P2₁/n space group) is shown to adopt a partial 1:1 ordering between cations in the B site whereas in the case of [CaLn₂]_A[CuTi₂]_BO₉ (Ln = Pr, Nd, Sm) and [CaLn₂]_A[ZnTi₂]_BO₉ (Ln = La, Pr, Nd, Eu) (Pbnm space group) the cationic distribution within the A and B sites is statistical [18,19,20]. In this work, we present the synthesis and structural properties of the SrLa₂FeCuSbO₉ compound, and their cationic distribution is obtained by using different models in the Rietveld refinement. We also present a study of the magnetic properties using magnetization measurements. The detailed relationship between the structural and magnetic properties was also investigated.

2 Experimental

The polycrystalline sample of SrLa₂FeCuSbO₉ was prepared by conventional solid-state reaction at high temperature from the mixture, in stoichiometric proportions, of La₂O₃, SrCO₃, Sb₂O₃, CuO, and Fe₂O₃ powders (purity > 99.95%). The mixture was ground, pressed, and annealed at 900 °C for 24 h, 1100 °C for 48 h, and 1200 °C for 48 h. The obtained sample is finally ground, pressed, and sintered at 1300 °C for 48 h. X-ray diffraction (XRD) experiment was performed at room temperature by means of a diffractometer Panalytical X′Pert Pro (θ-2θ) equipped with detector x’celerator radiance (Cu_Kα = 1.5406 Å) (45 kV, 40 mA), slope of divergence of 1°, 0.10 mm receiving slot, and anti-dispersion slot of 1°. Data were collected from 10 to 110° (2θ), in increments of 0.02°, with a count time of 15 s per step. The refinement of the crystal structure was performed by the Rietveld method [21] using the FullProf program [22]. The changes of magnetization as function of temperature (2–300 K) through zero-field cooling (ZFC) and field cooling (FC) procedures and under a fixed value of applied magnetic field of 50 Oe were performed using the Quantum Design MPMS-XL5 SQUID magnetometer. Magnetization measurements as a function of applied magnetic field (hysteresis loops) were recorded at 10 K from − 50 kOe up to + 50 kOe.

3 Results and Discussion

3.1 Structural Study

Obtained XRD pattern indicates that SrLa₂FeCuSbO₉ crystallizes in monoclinic symmetry (P2₁/n space group). The pattern also indicates the presence of a secondary phase with a small amount (about 7% by weight) which is attributed to La₃SbO₇ (JCPDS PDF#23-1138). The structural refinements were carried out by the Rietveld method using the FullProf program. The most intense peak region due to La₃SbO₇ was excluded from the refinement. The structural model used in our refinements was based on those reported for SrLa₂FeCoSbO₉ (P2₁/n space group) [10]. Note that in SrLa₂FeCoSbO₉, 1/3 of the Sr²⁺ ions and 2/3 of the La³⁺ ions are statistically distributed in the (4e) positions of A-perovskite sites. This structure contains two distinct (2c) and (2d) positions of the octahedral B-sites, which must be occupied by the smallest (Fe, Co, and Sb) elements. The oxygen ions are located in three different (4e) positions.

The first step of refinement was to determine the correct cell parameters using the P2₁/n space group. In structural refinement, three models were tested and the quality of the refinements was judged based on the reliability factors (R_P, R_wp, R_B, R_F-factor, and Chi²). The refined parameters were the lattice parameters, the oxygen atomic coordinates, the occupancy of the different sites, and the atomic displacement parameters of all atoms. The peak shapes were described using a pseudo-Voigt function.

The three tested models for the SrLa₂FeCuSbO₉ compound can be illustrated by the three crystallographic formula: SrLa₂(CuSb_0.5)_2c(FeSb_0.5)_2dO₉ in the first model (M1), SrLa₂(CuFe_0.5)_2c(SbFe_0.5)_2dO₉ in the second model (M2), and SrLa₂(SbCu_0.5)_2c(FeCu_0.5)_2dO₉ in the third model (M3). Figure 1 represents the observed and calculated shape of the selected peak region which is more sensitive to the cationic distribution, for the three models.

Fig. 1

Experimental (●●●) and calculated () XRD patterns, in the 2θ = 18–26° range of SrLa₂FeCuSbO₉, in the three models of cationic distribution

The M1 model, where the Sb⁵⁺ ions are distributed between the two (2c) and (2d) sites, gives unsatisfactory refinement results, in good agreement with reported work for the isotypic SrLa₂FeCoSbO₉ compound [10]. According to obtained reliability factors, the two M2 and M3 models seem to be the best. However, for the M2 model, where the Fe³⁺ ions are distributed between the two sites, the refinement leads to negative values for the parameters of displacement of oxygen and cations in the (2c) and (2d) sites. The M3 model, where the Cu²⁺ ions are distributed between the two (2c) and (2d) sites, gives satisfactory refinement results (R_B = 3.5%). Figure 2 shows the good agreement between the observed and calculated profiles obtained for the SrLa₂(SbCu_0.5)_2c(FeCu_0.5)_2dO₉ crystallographic formula. Table 1 displays details of the crystal structures, cell parameters, oxygen atomic coordinates, cationic distribution, and reliability factors.

Fig. 2

Experimental (●●●), calculated () and difference () profile of the XRD pattern of SrLa₂FeCuSbO₉ (*La₃SbO₇)

Table 1 Crystallographic data obtained from the Rietveld refinement for SrLa₂CuSbFeO₉

SrLa₂FeCuSbO₉ adopts a crystal structure with monoclinic P2₁/n space group. In this structure, 1/3 of the Sr²⁺ ions and 2/3 of the La³⁺ are statistically distributed in the A-perovskite sites. The two distinct (2c) and (2d) octahedral sites are occupied by (2/3Sb⁵⁺, 1/3Cu²⁺) and (2/3Fe³⁺, 1/3Cu²⁺), respectively. Table 2 groups the (La/Sr)–O, (Sb/Cu)–O and (Fe/Cu)–O bond distances and the different bond angles, and Fig. 3 shows the crystal structure of SrLa₂FeCuSbO₉. The way in which the octahedra are connected indicates that the crystal structure is typical of a perovskite with octahedral tilt distortion and that the tilt system is a⁺a⁺c⁻ [23]. We also note that the angles in the (Sb/Cu)O₆ and (Fe/Cu)O₆ octahedra are all far from those of the ideal structure of perovskite. This can be explained by the tolerance factor of Goldschmidt [24]. Taking into account the ionic distribution obtained by the Rietveld refinement and based on the effective Shannon’s ionic radii [25] (r(O²⁻) = 1.26 Å, r_VI(Cu²⁺) = 0.87 Å, r_VI(Fe³⁺) = 0.785 Å, r(La³⁺) = 1.356 Å, r_IV(Sb⁵⁺) = 0.74 Å, r(Sr²⁺) = 1.45 Å), the value of the calculated tolerance factor of Goldschmidt (t) is t = 0.90. This last value is low enough to give rise to significant deviations from the ideal perovskite. Note that the presence of Cu²⁺(3d⁹) ions in the two (2c) and (2d) sites can reinforce the distortion of octahedra by Jahn–Teller effect, but the amount of Cu²⁺ in every B site is not in favor of their appearance. The comparison of the experimental A–O and B–O band distance values with those calculated from Shannon’s ionic radii (i.e., (Sr/La)–O = 2.64 Å, (Sb/Cu)–O = 2.04 Å, and (Fe/Cu)–O = 2.09 Å) agrees well with the obtained cationic distribution (Table 2).

Table 2 Bond lengths (Å), bond angles (°), and bond valence sum (BVS) for SrLa₂FeCuSbO₉

Fig. 3

Crystal structure of SrLa₂FeCuSbO₉. a Cell in perspective. b Cell projected in (001) direction, putting to see the structure of the tilt system (a⁺a⁺c⁻). Sr²⁺ and La³⁺ in the A site are represented by a circle in green

The bond valence model [26] can give an estimate of the valences of cations and anions in a crystal structure. The value of calculated bond valence sum for each (Sb/Cu)_2c and (Fe/Cu)_2d sites are shown in Table 2. Note that a relativity good agreement between expected and calculated bond valence sums for the two octahedral (2c) and (2d) sites is obtained. The slight divergence observed for the (Sr/La) site, can be related to the disordered distribution of the two atom types within the same crystallographic position.

3.2 Magnetic Study

Figure 4 a shows the temperature dependence of zero field cooled (ZFC) and field cooled (FC) magnetization for SrLa₂FeCuSbO₉ recorded in an applied magnetic field of 50 Oe in the temperature range of (2–300) K. A clear transition from the paramagnetic to ferrimagnetic state is observed at T_C. The magnetic transition temperature T_C is extracted from the derivative dM(T)/dT (see Fig. 4b); both ZFC and FC magnetization curves show the same transition temperature T_C = 215 K. Fig. 5 displays the inverse magnetic susceptibility in the temperature ranges from 2 to 300 K. At high temperature (T > T_C), the inverse magnetic susceptibility is linear and can be described using Curie–Weiss law:

$$ \chi =\frac{C}{T-\theta } $$

Fig. 4

a ZFC/FC magnetization as a function of temperature. b Evolution of the derivative magnetization with respect to temperature

Fig. 5

Inverse of magnetic susceptibility χ⁻¹ as a function of temperature recorded under a magnetic field of 50 Oe. The red line shows the Curie–Weiss fit at high temperatures

Here, C is the Curie constant and θ is the Curie–Weiss temperature. The results of the fit using Curie–Weiss law are plotted in Fig. 5 with a solid line. The obtained Curie constant C and Weiss constant θ are 3.81 emu K/mol and + 226 K, respectively. Experimental effective magnetic moment (μ_eff) was extracted from the obtained value of the Curie constant (μ_eff = 5.51 μ_B/f.u). μ_eff was compared with the theoretical moment obtained using spin-only approximation for both Cu²⁺ and Fe³⁺ ions. The magnetic moments μ_S (Cu²⁺) and μ_S (Fe³⁺) can be expressed as ${\mu}_{\mathrm{s}}\left({M}^{n+}\right)=g\sqrt{S\left(S+1\right)}$, with g (g = 2) is the Landé factor, S (Cu²⁺) = 1/2 and S (Fe³⁺) = 5/2, which give μ_S (Cu²⁺) = 1.73 μ_B and μ_S (Fe³⁺) = 5.91 μ_B. The total theoretical moment was calculated using the following formula: ${\mu}_{\mathrm{theo}}=\sqrt{\mu_s^2\left({\mathrm{Cu}}^{2+}\right)+{\mu}_s^2\left({\mathrm{Fe}}^{3+}\right)}$ (μ_B/formula unit), the obtained theoretical moment is μ_theo = 6.15 μ_B, which is slightly higher than the experimental one (μ_eff = 5.51 μ_B) with a difference Δμ = |μ_eff-μ_theo| of 10.4%. This difference could be attributed to the presence of impurities (La₃SbO₇) detectable by XRD. On the other hand, the Curie–Weiss temperature value (θ = + 226 K) is high and positive which indicates the dominance of strong ferromagnetic interactions at low temperature. In addition, the transition temperature T_C = 215 K is very close to the Curie–Weiss temperature θ with a ratio of f = θ/T_C = 1.05. This obtained value of the f parameter is characteristic of a ferromagnetic behavior [27]. At low temperature (T < T_C), the magnetization curves ZFC and FC increase rapidly and stabilize at low temperature (see Fig. 4a), indicating the establishment of a ferrimagnetic order. The magnetization curves ZFC and FC show the existence of irreversibility over a wide temperature range (T < T_C). A similar difference in ZFC and FC magnetization data has been reported in disordered double and triple perovskite compounds La₃Ni_2-xCu_xB′O₉ (B′ = Sb, Ta, Nb) [28], La₂BMnO₆ (B = Ni or Co) [29]. This behavior has been explained by a disordered magnetic system such as spin glasses and clusters glasses [9, 12].

Figure 6 displays the magnetization as a function of the applied magnetic field at 10 K. The magnetization curve shows a weak hysteresis loop typical of a ferrimagnetic system with an unsaturated character up to 50 kOe, weak coercive field (H_C = 195 Oe) and remanent magnetization M_r = 0.45 μ_B/f.u. (see inset in Fig. 6). The first magnetization as a function of the magnetic field shows a rapid increase for low magnetic fields, followed by a linear increase as a function of the applied magnetic field without saturation, which indicates blocked magnetic moments due to the inhomogeneity in the distribution of magnetic moments caused by cationic disorder. The spontaneous magnetization M_s was deduced by extrapolating to H = 0 Oe the linear part of the magnetization curve; M_s is found to be close to 1.46 μ_B/f.u.

Fig. 6

Magnetization curves as a function of magnetic field up to 50 kOe recorded at a temperature of 10 K. Inset: zoom in the range of the magnetic field between − 3.5 kOe and + 3.5 kOe

It is of interest to note that the M_r/M_s ratio can give an idea on the structure of magnetic domains in ferromagnetic or ferrimagnetic materials. As well, it is known that for polycrystalline samples with uniaxial and cubic magnetocrystalline anisotropy, the M_r/M_s ratios are 0.50 and 0.83, respectively [30, 31]. Furthermore, the M_r/M_s ratio equal to or greater than 0.5 indicates that the sample can be in the form of nanoparticles with a single magnetic domain, and for M_r/M_s less than 0.5 it indicates the formation of a multidomain structure [32, 33]. In our case, the calculated M_r/M_s ratio, at T = 10 K, is approximately 0.30. This last value is less than 0.50, which could indicate the formation of a multidomain structure, in accordance with the structure of our compound as a bulk material. Otherwise, this low value of M_r/M_s can also be related to the inhomogeneity in the distribution of magnetic moments caused by the cationic disorder.

Assuming a G-type collinear ferrimagnetic arrangement, where Fe³⁺(S = 5/2), Cu²⁺(S = 1/2), and Sb⁵⁺ (diamagnetic) ions are distributed in a disordered manner in the two different (2c) and (2d) crystallographic sites, the spontaneous magnetization M_s can be estimated by the Néel model assuming a mean magnetization in each site. Considering the cationic distribution formula, the mean values of magnetizations M(2c) in the (2c) and M(2d) in the (2d) sites are 3.66 μ_B and 0.33 μ_B, respectively. Therefore, the net magnetization M_s can be calculated by the Néel model M_s = M(2c) − M(2d). The magnitude of the net magnetization M_s per formula unit should be 3.33 μ_B/f.u. However, the experimental value of 1.46 μ_B/f.u. is considerably less than the theoretical value, which indicates that the compound is far from saturated. This behavior has already been observed on triple-perovskite compounds [9, 15, 34, 35] which might be attributed to the following reasons. First, the cationic disorder imposed by non-magnetic ions such as Sb⁵⁺ ions disrupts the magnetic interactions, which prevents the establishment of a long-range magnetic order and leads to the formation of magnetic domains. Therefore, the low magnetization measured at low temperature is attributed to the spontaneous magnetization present in each domain. This behavior was observed on the partially disordered La₃Ni₂SbO₉ compound, which behaves as a ferrimagnetic relaxant below 110 K [9]. Secondly, the weak magnetization measured at low temperature is linked to the weakening of the efficiency and the dominance of the magnetic NN interactions by the presence of the diamagnetic Sb⁵⁺ ions. NN interactions are magnetic superexchange intercations between the nearest neighboring cations [10]. In this case, the long-range magnetic order is maintained, but the NN coupling is weakened. This is the case for the SrLa₂FeCoSbO₉ compound which has a magnetic transition temperature (T_C = 215 K) higher than that of La₃Ni₂SbO₉ [9, 10].

This second model seems to give an explanation of the magnetic behavior observed at low temperature in our compound which presents a cationic disorder with a relative high transition temperature (T_C = 215 K) similar to that observed for SrLa₂FeCoSbO₉. The positive Curie–Weiss temperature θ = +229 K, the shape of the ZFC/FC magnetization curves and the presence of a hysteresis loop are characteristic of a ferromagnetic behavior. However, the existence of a high irreversibility at low temperatures between the ZFC and FC magnetization curves and the low value of the spontaneous magnetization at H = 50 kOe indicates the presence of inhomogeneity in magnetic interactions caused by the cationic disorder. In order to improve the understanding of the magnetic interactions in this compound, neutron diffraction and electronic microscopy will be useful.

4 Conclusion

In summary, we have investigated the structural and magnetic properties of SrLa₂FeCuSbO₉. The crystal structure was refined by X-ray powder diffraction data in the monoclinic P2₁/n space group. The A sites are statistically occupied by 1/3 of Sr²⁺ ions and 2/3 of La³⁺. The two distinct octahedral (2c) and (2d) sites are occupied by (2/3Sb⁵⁺, 1/3Cu²⁺) and (2/3Fe³⁺, 1/3Cu²⁺), respectively. The result of the structural analysis indicates that the crystal structure of our compound is typical of a perovskite with octahedral tilt distortion and that the tilt system is a⁺a⁺c⁻. Magnetization measurements indicate a paramagnetic–ferrimagnetic transition around T_C = 215 K. In the paramagnetic region, the magnetic susceptibility was described by Curie–Weiss law. The experimental effective moments were explained by only spin contribution for the Fe³⁺ and Cu²⁺ ions. The effective magnetic moment obtained experimentally is μ_eff = 5.51 μ_B/f.u. The positive and high Curie–Weiss temperature (θ = +226 K) indicates the dominance of strong ferromagnetic interactions at low temperature. The obtained value of the frustration parameter is characteristic of a ferromagnetic behavior. In the ordered magnetic regime, the shape of the ZFC/FC magnetization curves and the presence of a hysteresis loop are characteristic of a ferrimagnetic behavior. However, the existence of a high irreversibility at low temperatures between the ZFC and FC magnetization curves and the low value of the spontaneous magnetization at H = 50 kOe indicates the presence of inhomogeneity in magnetic interactions caused by the cationic disorder. Finally, these results will be useful to understand the magnetic properties in SrLa₂FeCuSbO₉.